Smart technology for a strong manufacturing industry with 24/7 uninterrupted production of IIoT solutions for companies.
Until the seventies and early eighties of the previous century, measuring equipment and related data transmission was mostly analog.
An analog signal is a translation of a physical property to an electrical property over a defined range. E.g a pt100 temperature sensor translates temperature to electrical resistance. Between 0°C and 100°C the resistance will change between 100Ω and 138,5Ω.
On the gauge on which the measured temperature is read, a scale in °C will be present, but technically the gauge is indicating the resistance of the sensor.
It is obvious that the length of the wires between the sensor and the readout affects the readout of the gauge. Indeed, 3 or 4 wire connections, combined with some clever compensation circuitry, may theoretically compensate for the wire lengths, but even then connections and the readout itself can introduce variations which you can only detect and adjust for by calibrating the system in place.
The same holds true when the resistance is converted to more robust properties as voltage or current.
And given the fact that all elements in the measuring loop are subject to slow deterioration, regular calibration is a must to ensure date integrity on the long term.
With the introduction of DCS and PLC systems, the analog signals needed to be converted into digital signals before the automation system could use them in its internal logic. The circuitry which does this conversion is called an ADC or an analog to digital converter. For converting a digital signal to an analog signal a DAC or digital to analog converter is used.
A DAC converts the analog signals into a sequence of two possible values, representing zeros or ones. Once this conversion is done, a lot of new possibilities to screw up the measured data become available.
The raw output of the DAC must be scaled to meaningful values using scaling factors which can be wrong.
When data is exchanged between systems, conversion errors can be introduced if different representations are assumed (e.g. signed or unsigned).
When data is transmitted over a long distance the zero and one values can be distorted to such an extent that zeros are interpreted as ones and vice versa.
The software can contain bugs resulting in values being swapped.
In other words: digitizing your data does not mean data integrity is ensured. But calibration is not the solution to ensure data integrity.
Only the analog part of a measuring loop is subject to calibration, the data streams after the ADC is subject to computer systems validation. Digital data transmissions should include error detection algorithms, and security measures against viruses and intruders should be present.
Although computer systems validation is far more complex and elaborate as calibration, there is no need to repeat it every year if adequate change control and performance monitoring procedures are in place.
With electronics becoming cheaper and smaller for many decades, sensors with build in ADC have become available for almost all physical properties. Data are then transmitted via a digital communication protocol like modbus, OPC UA or BACnet.
As you can see, the part subject to calibration is fully contained in the sensor itself. When calibration of the sensor is due, there is no problem in swapping it out with a freshly calibrated sensor, provided the digital communication is exactly the same. To ensure the latter, a procedure to set the address and communication parameters of the new sensors to the same values as the one being replaced may be necessary. But in situ calibration is no longer a requirement.
Fun fact: some sensors have a build in ADC to digitize the actual measured value at first, after which a build in DAC converts it back to an analog 4-20mA or 0-10V signal, depending on the configuration you select via digital(!) communication. If used like this, in situ calibration remains a requirement.
